TL;DR: Ollama v0.40.4 introduces direct command-line support for decision models, enabling quick classification and yes/no queries from the terminal.
Summary: Ollama v0.40.4 now allows users to run decision models directly from the terminal using ollama run. This feature supports tasks like asking yes/no questions, classifying text into labels, or scoring against a rubric, with examples provided for ticket analysis. Decision models are also supported by the MLX runner on Apple silicon and llama.cpp.
Why it matters: This update simplifies integrating AI-powered classification and decision-making into command-line workflows and scripts. AI builders should explore using this for rapid prototyping of text classification, sentiment analysis, or automated triage directly within their development environment.
Source: github_releases